How Many Trains Will Fit Through Union Station?

During the Metrolinx Electrification Study, those of us who attended various workshops became aware that there was a parallel study of capacity issues at Union Station.  The electrification plans are, among other things, in support of operating better service on GO generally, but if that service won’t physically fit through Union Station and its approach corridors, there’s a big problem.

That problem is independent of electrification per se because The Big Move from Metrolinx depends on substantially improved commuter rail service.  No capacity, no additional service.

At the recent Metrolinx Board meeting, GO’s President, Gary McNeil, presented an update on GO operations and construction activity.

GO President’s Report & Presentation Deck

The report includes a reference to Union Station capacity:

… Retaining wall construction is well underway to allow for an additional track in this corridor. The Union Station capacity study has been completed, with the result that in the near term, there is capacity at this station to meet needs. With the start of design of double berthing and new south platform, this will provide access required for service expansions. This work is anticipated to be completed in the next five years.  [Page 8]

After the meeting, I requested a copy of the study to learn what conclusions it might have reached. Various working papers from the study had been leaked, but they were neither definitive nor entirely coherent on how to deal with the problem.

Metrolinx has now replied that:

At this time, a detailed public component of the Union Station study is premature as we are undertaking on-going research. Specific information will most likely be available for the public when future potential projects develop from this study.

The purpose of the study is to assess the Union Station Rail Corridor (USRC) train capacity at four time points:

  • existing;
  • completion of planned infrastructure in 2015 and implementation for service improvements, including the ARL;
  • Electrification Reference Case (ERC);
  • and 2031 (Big Move planning document).

In doing so, we hope to identify opportunities to increase capacity by making more effective use of existing and planned infrastructure.  We also hope to identify the infrastructure needed to address any capacity shortfalls.  This study provided only a technical analysis, and Metrolinx will consider its opportunities after further assessment.

However, there is a good deal of material to get started on.

First off, the capacity problem is not some far-distant issue, but one that GO must address today.

  • Current schedules and operating patterns use all available capacity in peak times (24 GO Trains per hour in peak time). Bottlenecks are happening in the Union Station Rail Corridor (USRC), not in the Union Station train shed and there cannot be any additional peak trains without negatively affecting the on-time performance.
  • By 2015, there will 24 GO Trains per hour in peak time, plus 4 Air Rail Link trains. At this time, service levels can be accommodated with the planned infrastructure.

This means that if you were hoping for additional peak hour service in the next four years, you shouldn’t hold your breath.  Existing runs may be extended to more remote destinations, but there isn’t capacity for more trains in that peak hour.  This also means that the number of commuters entering Union is capped by what these trains can carry.  Extension to 12-car trains will help, but that is of limited benefit on corridors that already have long trains.

Looking at the higher service level in the Electrification Reference Case, the study concluded that:

  • the service did not operate with an acceptable on-time performance level (based on the RTP/GO 2020 planning documents).
  • the ERC provided a conceptual service level, not an operational schedule.
  • with minor adjustments, an actual ERC schedule could function with an acceptable on-time performance.
  • [the service] requires a new south platform and changes to both GO and VIA operations.
  • up to 44 GO Trains per hour in the peak is the approximate capacity upper limit.

The Big Move requires an even higher level of service on the GO corridors than in the Electrification Base case, and this brings more challenges.

  • significant capacity shortfall inside and outside of the USRC
  • additional tracks under Union would need to be installed (4 tracks required). Entraces to the underground would likely be problematic as extra property is required in developed areas. The minimum level below the existing track level would be 23.5 metres and the vertical access requirement would conflict with the City retail area.

Something not mentioned here, but definitely a problem for any construction, is that the ramps to access new underground platforms would have to start well east and west of the station in order to be low enough once they reached the heart of the rail corridor.  There is also the small matter of building under a working corridor already at capacity, and the high water table.

It is unclear whether any allowance was made in the study for the different performance characteristics of electric versus diesel trains.

Other findings from the study include:

  • Turning trains (versus through trains) increases track occupancy time by 56%, reducing capacity.
  • With double berthing, turning trains allows 8 trains per hour compared to 6 through trains per hour, which does increase capacity.
  • Wider platforms achieved by track removal is found to reduce train capacity.

Recommendations are:

  • Develop a new south platform by removal of Track 16 (planning and design work is underway).
  • Improve capacity (and reliability) with infrastructure modifications to increase straight routing.
  • Make relatively minor infrastructure changes to improve the utility of the two connecting tracks for commuter service, while protecting for freight operation.
  • Change the GO platform assignments (by 2015) to make better use of capacity.
  • Investigate ways to increase platform use with VIA (conversation is underway).
  • Proceed with double berthing (work is underway).

Little of this is much surprise to those who have watched and cared for Union Station.  With the disappearance of the rail yards and the construction of the CN Tower nearly 40 years ago, preserving capacity in the rail corridor was not high on the priority list.  The railways still owned the corridor, and their interest was to maximize development potential.

GO Transit itself was less than 10 years old, and the idea that Union Station would someday need to accommodate vastly more trains and passengers was not on their radar.  Only the Lakeshore route operated until Georgetown joined the network in 1974 (the year the CN Tower was topped off), and the double-deck coaches didn’t come into service until 1979, the year GO moved into its “new” east concourse.

Growth in commuting to downtown has been almost entirely handled by GO over past decades while demand on both streets and transit was limited by capacity and parking supply.  But even GO has its constraints.  While much attention has focused on parking lots and garages all over the GTA, the capacity at Union wasn’t on the table.

With the station’s reconstruction, now underway, GO will get over double the concourse space of the old east wing station, but the trains will require new approaches to operations.  Metrolinx planning can no longer afford to assume unlimited capacity in the rail network (or the transit network for that matter).  The Big Move 2.0 must include plans for new local and regional infrastructure, not just the movement of virtual passengers in a demand model.

Downtown Toronto continues to grow and it will remain the centre of jobs and office space for decades to come.  Yes, there are new schemes in the outer suburbs, but any of these would fit comfortably in the core area and vanish among what is already there.  If nothing else, the growth of a “Metropolitan Centre” beyond downtown Toronto simply cannot occur because there isn’t enough transportation capacity in one place to support such a scheme.

Future growth in downtown will depend on having more people living closer to where they work rather than a train ride away in the 905.  This has already begun with condos in walking or short transit distance of the core, and will spread to the eastern waterfront.  However, if transit doesn’t increase to match this demand, the TTC (and Metrolinx for that matter) will have failed.

Metrolinx crows about its importance in reducing commute trips, pollution and energy needs, but ignores the most important asset right outside its doors — the local transit system.  Toronto scrounges for money wherever it can be found, and gets distracted by dreams of new subways and dubious financing schemes.

Just as the expressway, road and parking networks have filled, and today’s challenge is how to wean drivers onto transit, the transit system itself has constraints, some physical, some financial.  We are reaping the joys of suburban expansion, but it’s time to take capacity into downtown Toronto much more seriously.

49 thoughts on “How Many Trains Will Fit Through Union Station?

  1. Karl said: “You’re suggesting Windsor-Oakville trains”…

    No. I am suggesting that terminating services CONTINUE beyond Toronto and turn back rather than terminating at Union. I am therefore suggesting Windsor-Oshawa, Kingston-Aldershot/Niagara Falls and so on.

    Steve: Several trains that appear to terminate at Union actually run through, although usually after a non-trivial layover. For example, the morning train from Stratford becomes an Ottawa train.

    Those that do end at Union don’t stay there, but run through to the Via shops at Mimico. Where the train is going after it leaves Union is not the problem. The long platform dwell time is.


  2. Steve: Several trains that appear to terminate at Union actually run through, although usually after a non-trivial layover. For example, the morning train from Stratford becomes an Ottawa train.

    That specific example is no longer correct. There are two morning arrivals that run through Stratford: #86 which originates in London and #84 which originates in Sarnia. Both now go to the VIA yard in Mimico (really in New Toronto, but everybody calls it Mimico) after arrival at Union Station.

    There are other VIA trains that behave in the way Steve says. #90 from Niagara Falls becomes #54 to Montreal via Ottawa after a 1 h layover at Union Station. #70 from Windsor becomes #60 to Montreal after a 1 h 31 min layover. #57 from Montreal becomes #75 to Windsor after a 1 h 23 min layover. #72 from Windsor becomes #46 to Ottawa after a 2 h 19 min layover. #61 from Montreal becomes #79 to Windsor after a 2 h 20 min layover.

    VIA recently changed its schedules, greatly reducing the dwell time in Ottawa for equipment running through that station from Montreal to Toronto or vice versa. I don’t know if they’ve ever considered doing the same thing in Toronto.

    Steve: Thanks for the update.


  3. Brainstormed an idea: Maybe GO could try and find ways to divert some trips on to the Bloor Danforth line? I predict that a lot of the passengers which currently use it will move to the Crosstown line once completed, so capacity might be less of an issue than it is now. Using Exhibition Station could work as well.

    Would it be possible to turn trains around at Bloor, Exhibition, and Danforth Stations? If not now, what about with infrastructure improvements?

    Steve: I will believe that there will be surplus capacity on the BD line when I actually see it. Although the Crosstown line will draw some traffic north, the closer one gets to Yonge Street, the more the existing BD demand will stay where it is. To what extent the demand on Eglinton is made up of net new riders (who might otherwise have gone via BD) as opposed to diversions of existing ridership is hard to say. There are other variables (such as a DRL) in the mix too.

    Diverting traffic from GO onto the BD line chews up relatively expensive subway capacity and is not necessarily the best use of this infrastructure. To the degree that these travellers are bound for the core area, they are still going to be dumped onto the south end of the Yonge line where today they are riding counter peak north from Union.


  4. so if we’re cataloging crazy ideas. What about building a line through the harbour and reroute lakeshore line or via through it. It could be done easily with some prefab tunnels sunk into the lake. Go from the ex to dvp. With some dredging to keep it deep enough. Potentially with a station on the waterfront with a connection to the island airport. Maybe add another station near the Portlands then join back at dvp.

    Steve: Whatever you are smoking or drinking or otherwise consuming, I think there is an untapped market. However, you will have to flog it on your own site.

    Or maybe it’s the heat.


  5. You may not have noticed this but GO is in the process of replacing the throats of Union Station at both ends, including replacing all the switches and the signalling system. (I’m curious where they’re putting their new dispatching center, as they appear to be putting TTR out of business.) This is supposed to include faster switches (no more 25 km/hr speed limits). They have also specified that platforms will be reallocated.

    This is probably how they’re planning to handle the short-term capacity problems. The question becomes how much they can get out of a faster throat and better platform allocation.

    Steve: Yes, this work has been underway for some time, but GO’s capacity at Union, even with all the changes, does not meet projected needs of future services. This begs a question: to what degree were the changes now in progress co-ordinated with service plans. One side of the organization loves to announce new services, but another side needs to be able to physically operate them. Until fairly recently, GO, Metrolinx and Queen’s Park operated more or less independently on planning and announcing how the network would grow. Indeed, The Big Move completely ignored capacity constraints on the growth of riding on both the GO and TTC rail networks.


  6. Hmm. Thinking about it, current projects (mostly “Georgetown South”) will have tracks coming in from the west as follows (listed north to south)

    2 tracks Barrie
    4 tracks Georgetown (/Pearson/VIA)
    2 tracks Milton
    ??? tracks Lakeshore (/VIA)

    There are actually enough approach tracks for every line to come in “straight” without merges.

    Not taking account of the “underpass”. East of the “underpass”, you have

    1 through track (with room for expansion)
    3 tracks underpass (from the tracks further south, Lakeshore or Milton)
    2 tracks overpass (from the tracks further north, everything but Lakeshore)
    2 through tracks Lakeshore

    So, OK, there are almost enough tracks to avoid merges.

    And from the east as follows (listed north to south)

    2 tracks Richmond Hill (/ VIA / Ontario Northland)
    2 tracks Stouffville
    2 tracks Lakeshore East (/VIA)

    This is complicated by the fact that there aren’t actually six tracks on the approach on this end.

    I’m ignoring freight for obvious reasons; not enough to worry about.

    The traffic is unbalanced, with more coming from the west than from the east. But it doesn’t line up that badly. The trains which need really long dwells, the Canadian and the Ontario Northland, can be lined up on the eastern half of some platforms, opposite the “excess” GO Transit trains from the west side. The remaining platforms can be devoted to trains running through from the west side to the east side, and vice versa, fully directional, including VIA Corridor trains. There’s plenty of space; I’d expect the remaining chokepoint to be east of Union, since it looks like the trains could be fully sorted moving west of Union, with no significant conflicts.

    VIA corridor trains would have to have much shorter dwell times, of course. (I guess GO might need shorter dwells too.) And more importantly, they’d have to start letting people hang out on the platforms, rather than going through the ludicrous line-up-single-file procedure which currently slows everything down.

    Steve: Another important point is that both VIA and GO have yard moves to/from the Mimico, Bathurst and Don yards. These have to be factored into capacity on the approaches. It’s ironic that CN and CP ditched their huge passenger yards to free up land for development, and decades later GO established new yards close to Union


  7. @Nathanael Nerode: Lakeshore needs more than two tracks to avoid conflicts due to its mix of local and express trains, in addition to the non-revenue movements Steve noted.


  8. Does GO store any trains at the platforms off-peak to avoid yard moves? If not, is this feasible/helpful?

    Steve: No and no. Only trains that came in at the very end of the am peak and went out at the start of the pm peak could be stored without blocking tracks, and this would contribute nothing to peak hour capacity.


  9. Was it ever envisioned that a DRL would connect with an Agincourt-Dupont GO line (so that the Agincourt GO line wouldn’t dump passengers on the Yonge line via Summerhill station)?

    If that is what is envisioned, then why have a DRL stop only at Eglinton, 500 metres short of the rail line?

    Steve: As and when an Agincourt GO line gets service, that’s an extension worth thinking of. What is so frustrating about the TTC is that they refuse to consider the line north of Danforth even though the Eglinton terminus was in the 1960s plan. Meanwhile, they wasted much time during the Transit City studies claiming that they could somehow fit a surface LRT onto narrow streets in East York.

    Liked by 1 person

Comments are closed.